0

first of all I'm a hobbyist so it's ok to roast me provided it comes with good insights!

I am prototyping an electronic apparatus comprised of different parts

  • a microcontroller running on 5V (actually 3.3 but it has its own voltage regulator 5V->3v3, negligible power draw);
  • a strip of 150 RGB 5V LEDs. Maximum current draw per LED at max white brightness appears to be around 60mA. Total is then 150 * 60 = 9Amps ==> 45Watt
  • a class D amplifier, 40W maximum at 12V, 20+20W (so I guess, maximum 3A roughly)

At the moment it's all a mess of 3 different power supplies and I wanted to simplify and streamline it.

I have a 12V10A (120W) DC supply, whose output would be split between the amp at 12V and a 300W buck converter to create the 5V.

I am trying to figure out how much current can I actually safely draw from the PSU, as I don't really know how the buck converter would behave in such a circumstance or how to approach it.

if I only had 12V loads, the answer would be easy - I can draw maximum 10Amps. But I have 12V loads and 5V loads and if I sum the maximum power and current requirements of every single component, I am well below the 120W BUT I go above those 10 amps.

What is the correct way of looking at it to calculate the maximum current draw / maximum current available to the amps and the LED strip?

Power? i.e. as long as the combined power draw of the amp and buck converter powering the strip and MCU is under 120W, it won't burn my house down even though the sum of the currents is above the maximum draw for the psu because the buck converter converts the power into lower voltage THUS higher available amps, even above the 10A by itself?

or Current? i.e. I need to make sure that the combined current draw of all components does not exceed the PSU's 10 amps, no matter the voltage?

Thanks to anyone who makes sure I don't burn a hole in the floor with mad empirical experimentation.

RedSash
  • 11
  • 1
  • Here's a helpful reference that you should review if you haven't already, about sizing power supplies and loads: https://electronics.stackexchange.com/q/34745/2028 – JYelton May 18 '22 at 17:04
  • 2
    It depends on where you bought the converter. It may genuinely go to 20A as advertised, or if the seller is unscrupulous, they advertised 20A, but in practice it will overheat at 10A for any reasonable ambient temperature. Does your converter have a datasheet? Did you buy it on AliExpress? – Math Keeps Me Busy May 18 '22 at 17:04
  • How much fire are you willing to put out? – Reinderien May 18 '22 at 17:14
  • you should be fine because your LEDs and amp won't hardly ever draw full current. – dandavis May 18 '22 at 18:34
  • Have you considered using 12v RGB LEDs? Then you could eliminate the buck converter entirely. – user1850479 May 18 '22 at 18:52
  • This. This is why in AC mains, when tallying up various different-voltage system loads, we use VA for everything. @user1850479 most likely OP's LEDs are individually addressable. – Harper - Reinstate Monica May 19 '22 at 02:47
  • @Harper-ReinstateUkraine You can get 5 or 12v addressable LEDs. Seems strange to pick 5v LEDs if you have a 12v source. – user1850479 May 19 '22 at 02:52
  • @user1850479 Harper-ReinstateUkraine well, it's mostly driven by what I have at hand. Like 4x 5mt long strip of individually addressable ws2812 leds which work on 5v but also the fact that the microcontroller still needs a 3v3 or 5v supply. So leds are powered by 5v and driven by the 3v3 mcu pins via a 3v3 to 5v logic level switcher. Had I 12v leds, I'd still have to use a buck to power the mcu (even though I would have been less concerned about current and power draw from it) – RedSash May 19 '22 at 11:05
  • You can power an MCU using a 25 cent linear regulator, buck converter is probably overkill unless it's a very high power device. – user1850479 May 19 '22 at 11:21
  • @user1850479 no it"s an esp32, it should draw 1-1.25 W peak, most likely half a watt or less. I'm not savvy with either linear regulators nor buck converters. As it is supposed to be an always on device, what would be the dissipation and heat considerations for both cases? AFAIK linear regulators are a bit wasteful. (still made a note in my mind to consider designing with 12v leds in mind for next iterations) – RedSash May 19 '22 at 12:07
  • @user1850479 by the way I just checked on adafruit to see the current and power requirements of 12v led strips - still a maximum of 60 milliamp per led so 9a which brings it all to 84w. Yowza. I was hoping for maybe lower currents due to higher voltage but it seems it's going to be too power hungry to consider – RedSash May 19 '22 at 12:15
  • The 12v strips will use proportionally less current. The most efficient solution is to run the strips at the power supply voltage, but if you already have the strips and don't want to buy a different one you may just have to accept some loss. – user1850479 May 19 '22 at 12:23

1 Answers1

1

You have a 45 watt 5V load powered through a buck converter at let's be conservative and say 85% efficiency, which means 53W at the input. The amplifier is 40W so that's a total of 93W draw on your 12V power supply. The 12V supply is rated for 120W so you should be fine but you need to look at the datasheet and see if the 120W is continuous and if there's any derating that needs to be taken into account. Also, depending on the shall we say provenance of your 12V supply, it might be best to observe it closely at maximum system power for a bit before letting it out of your sight.

vir
  • 14,718
  • 13
  • 28
  • thanks. so I seem to understand that it is correct to think in terms of power (I. E. The sum of all individual currents can be well above the 10a - the buck could even supply 15a, provided that the total power is under 120w)? Is that the correct way to look at it? – RedSash May 19 '22 at 11:09